The Development of an Assessment To Evaluate Behavioral
Factors Associated With Falling
KEY WORDS
• accidental falls
• aged
• validity
Lindy Clemson, MAppSc(OT), is Lecturer, School of
Occupation & Leisure Sciences, Faculty of Health
Sciences, The University of Sydney, PO Box 170,
Lidcombe, 1825 Australia; L.Clemson@fhs.usyd.edu.au
Robert G. Cumming, PhD, is Professor, School of Public
Health, The University of Sydney, Lidcombe, Australia
and the Centre for Education and Research on Ageing,
Concord Hospital, Concord, Australia.
Dr. Robert Heard, PhD, is Senior Lecturer, School of
Behavioral Sciences, Faculty of Health Sciences,
The University of Sydney, Lidcombe, Australia.
380
Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930150/ on 07/09/2017 Terms of Use: http://AOTA.org/terms
Lindy Clemson,
Robert G. Cumming,
Robert Heard
The purpose of this study was to report the development of the Falls Behavioral (FaB) Scale for Older People,
an assessment designed to evaluate behavioral factors that could potentially protect against falling. Instrument
development included content analysis, expert review, and factor analysis. Ten behavioral dimensions were
identified including Cognitive Adaptations, Protective Mobility, Avoidance, Awareness, Pace, Practical
Strategies, Displacing Activities, Being Observant, Changes in Level, and Getting to the Phone. The final 30-
item scale had a Content Validity Index of 0.93. Test–retest reliability was ICC = 0.94 (p < .01). Construct valid-
ity was established by showing that, as expected, scale scores were positively associated with increasing age
(rs = 0.46, p < .01) and negatively associated with greater physical mobility (rs = -0.68, p < .01). People who
had fallen utilized safer behaviors than those who had not reported a fall (p < .05) providing a benchmark for
using the scale in future studies. The FaB is an easily completed, reliable, and valid tool for determining the
presence or absence of protective behaviors. It has potential to assist in goal setting for falls prevention and to
evaluate behavioral outcomes of fall prevention programs.
Clemson, L., Cumming, R. G., & Heard, R. (2003). The development of an assessment to evaluate behavioral factors asso-
ciated with falling. American Journal of Occupational Therapy, 57, 380–388.
alls are the leading cause of morbidity and mortality due to injury in elderly per-
F sons and are associated with substantial medical and rehabilitation costs, as well
as social isolation and premature institutionalization (Tinetti & Williams, 1997).
Occupational therapists can assist older people to reduce falls by home modifica-
tion and adaptation, at least with those who report they have fallen in the past
(Close et al., 1999; Cumming et al., 1999). An occupational therapy approach
encompasses both environmental change and the persons’ interaction with the
environment, their actions, and their behavioral adaptations at home and in the
community. Attitudinal factors about falls prevention and home safety have been
found to have a large impact on whether people adhere to recommendations for
home modification (Clemson, Cusick, & Fozzard, 1999; Cumming et al., 2001).
Behavioral as well as environmental assessment tools are needed in order to edu-
cate elderly people about strategies to reduce their risk of falling.
Known risk factors for falling include increasing age, history of falling, mobil-
ity impairment, and reduced lower limb strength, and behaviors associated with
falls are reduced activity and leaving home less often (e.g., Tinetti, Speechley, &
Ginter, 1988). A fall can be conceptualized as a sequential event with a number of
contributing personal, environmental, and behavioral factors (Cummings &
Nevitt, 1989). Viewing the events leading up to and surrounding the fall provides
additional information about the nature of the risks and the kinds of specific
sequences of actions that contribute to falls.
In studies presenting self-reports of causes of falls (e.g., Hill, Schwarz, Flicker,
& Carroll, 1999) participants frequently attribute actions such as “hurrying,”
July/August 2003, Volume 57/Number 4
 “carelessness,” and “inattention” as causal fall behaviors.
Connell and Wolf (1997) and, more recently, Clemson,
Manor, and Fitzgerald (in press) have examined the cir-
cumstances of falls from an in-depth qualitative perspective.
These latter studies used techniques to reconstruct fall
events and situate them within the contexts of both con-
tributing behaviors and environmental circumstances. The
researchers identified patterns of environmental-behavioral
factors related to fall events such as failing to avoid tempo-
rary hazards, excessive attention to other aspects of sur-
roundings or other intentions rather than paying attention
to the route ahead, and habitual and inappropriate envi-
ronmental use. In both studies the researchers describe
a dynamic relationship between the environment and
behavior.
Behaviors can be examined both in terms of the degree
of risk confronting the older person as well as targets for
behavioral change to more protective behavior (Gochman,
1997). Shiffrin and Schneider (1977) have made the dis-
tinction between habitual behaviors as automatic cognitive
processes and intentional behaviors as controlled cognitive
processes. Ronis, Yates, and Kirscht (1989) explain that
habits are set in motion by situational cues, such as a spe-
cific place and a specific time. For example, an older person
may always enter his or her home in the same habitual man-
ner by moving to the left side of the front steps and hold-
ing onto the left handrail. Intentional behaviors (Shiffrin &
Schneider) require reflection, intention, and planning. A
windy day may prompt reflection about the strength of the
wind and how this may challenge the person’s balance; the
reflection acts as a cue to intentionally decide how best to
prepare for the day, or even whether or not to go outside.
This decision requires some deliberation and planning and
there are alternate actions. Such actions are not automatic.
Knowing what these target behaviors are, whether habitual
or intentional, is the first step to utilizing a behavioral
change approach in falls prevention (Cole, Berger, &
Garrity, 1988). The close interaction of habitual and inten-
tional behaviors with situational and environmental cues
assists in delineating the domain of interest for an assess-
ment.
Some behaviors, actions, or decisions may be risky for
one person or be less risky or pose little risk for another per-
son. For example, not holding onto the left handrail may be
risky for one person or not risky for another. Holding onto
the left handrail could be described as protective for the per-
son who is at risk. So one would expect that some people
would engage in what could be protective behaviors more
often than others. Jackson (1996) recognized how elderly
people use adaptive strategies following a fall experience,
and how these adaptations occur as they balance resources
The American Journal of Occupational Therapy
Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930150/ on 07/09/2017 Terms of Use: http://AOTA.org/terms
and limitations against a background of personal meaning
connected to occupation. Occupational therapists work
with people to assist them in enhancing such adaptations
and in making safe adaptations. A behavioral assessment
may provide a tool to raise awareness for the therapist and
the older person of the kinds of subtle, everyday habitual or
intentional behaviors that occur.
The ability to measure the impact of fall prevention
approaches is essential if we are to successfully evaluate the
efficacy of those interventions that aim to facilitate behav-
ioral adaptations related to managing fall risk. Studies to
date have relied on measures of perceived health, perception
of risk, and measures of self-efficacy. No instruments have
been found that provide a broad behavioral assessment. The
objectives of this research were to develop a reliable and
valid assessment tool to evaluate the kinds of day-to-day
behaviors and actions that offer an older person protection
from falling during daily activity. The tool was limited to
the older person’s perceptions about their behaviors and
designed to be either self-administered or administered by
interview. This paper describes the development of the Falls
Behavioral (FaB) Scale for Older People1, a scale that meets
these criteria.
Design and Methods
Development of the FaB scale involved content analysis of
relevant literature and refinement by an expert review, fol-
lowed by a factor analysis. Factor analysis is an ideal method
to assist with test construction (Comrey & Lee, 1992;
Kline, 1994; Thurstone, 1947). Factor analysis is intended
to identify sets of items that respondents have rated. In this
way the larger number of original items is reduced to a
smaller number of sets called “factors,” each factor assessing
one common underlying dimension. This process identifies
the underlying content areas of the phenomenon of interest
and, hence, further contributes to the tool’s validity. Next, a
survey of older people assisted in providing information to
test the construct validity of the FaB by determining its rela-
tionship to falling and risk factors of falls. Finally, another
sample was used to determine the test–retest reliability of
the final 30-item FaB scale over a 2-week period. Ethics
approval to conduct the study was obtained from a univer-
sity Human Ethics Committee.
1 The instrument and manual (including a complete table of factor coef-
ficient loadings) can be downloaded from www.ot.fhs.usyd.edu.au/pub-
lications.html.
381
 Initial Instrument Development
To establish content validity, the FaB scale was initially
developed from a content analysis of literature and an
expert review process similar to procedures used by
Clemson, Fitzgerald, and Heard (1999). A content analysis
was conducted of all literature that could be located that
studied falls and provided some information on behavioral
causes of falling in older people. MEDLINE and CINAHL
databases were searched for the period spanning 1966 to
1999. Information from the content analysis of literature
was listed and operationalized into behavioral statements to
make up the initial 54-item FaB scale.
An expert review was conducted to confirm face valid-
ity of the scale. Lynn (1986) recommends using at least five
judges to provide a sufficient level of control for chance.
Nine judges were selected for their expertise in the area of
falls and to cover a broad perspective of related content,
including two lay persons to represent the views of older
people. One expert returned the review with comments
and did not complete the rating scale. The judges rated
each item of the initial 54-item using a four-point rating
scale (Ebel, 1979; Skakun & Kling, 1980): Essential,
Acceptable, Questionable, and Not relevant. The judges’
ratings were quantified using the content validity index
(CVI) proposed by Waltz and Bausell (1981) and adapted
by Lynn, which further guided item selection. Three items
that were variations of a question about bifocals were
retained as judges failed to discriminate between them.
The expert review resulted in the 45-item version ready for
the factor analysis.
The Participants and the Survey
Aged care organizations and seniors’ groups were
approached to assist with recruitment of people 65 years or
older living at home, excluding nursing homes and hospi-
tals. A member of each participating organization asked res-
idents or members of a group within the organization to
participate. Thus, 550 volunteers from a wide variety of
community organizations were invited to complete the 45-
item FaB scale plus a survey of health and demographic
information. Participants were given an explanatory letter
and a stamped self-addressed envelope for returning the FaB
scale and the survey by mail.
The FaB scale was a list of, at this stage, 45 behavioral
statements “describing things that we do in our everyday
lives” with multiple choice responses of never, sometimes,
often, always, or does not apply. A brief introductory para-
graph included a definition of a fall and instructions. Items
were scored from 1 (never) to 4 (always) with 0 for does not
apply. Prior to analysis, six items were reversed so that all the
382
Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930150/ on 07/09/2017 Terms of Use: http://AOTA.org/terms
high scores indicated the presence of protective behaviors
and low scores the absence of protection.
Demographic information collected included age, gen-
der, cultural and ethnic background (country of birth and
language mostly spoken at home), and type of dwelling.
Respondents were asked if they had had a fall in the past
year and, if so, how many falls. Another question addressed
the extent of community mobility (how many times left
home in the past month).
The Physical Functioning subcomponent of the SF-36
Health Survey (Ware, Snow, Kosinski, & Gandek, 1993)
was also administered to all participants. This scale indicates
the extent to which, on a typical day, a person’s health lim-
its them in performing a range of physical activities such as
carrying groceries, climbing stairs, bathing, and dressing.
High SF-36 mobility scores means better mobility. There is
evidence of validity for use of the SF-36 with elderly people
(Lyons, Perry, & Littlepage, 1994) and Australian norms are
available (Australian Bureau of Statistics, 1995) for com-
parison.
Statistical Analysis
Factor Analysis. All analysis was carried out using the SPSS
(Version 10.0, 1999) statistical program. Items were
screened for normality of distribution and the few non-nor-
mal items were either transformed or, if they had almost no
variation, dropped from the analysis. A principal compo-
nent factor analysis using varimax rotation with Kaiser nor-
malization was conducted to identify groups of related
items (Kaiser, 1958). This process obtains the optimum
number of factors using an eigenvalue > 1. Only items with
an eigenvalue > 1 were retained. We tried other extraction
and rotation methods, however, they all produced what we
regarded as inferior solutions. They (1) did not produce fac-
tors within 25 iterations or (2) accounted for less of the
variance or (3) produced factors on which many items did
not load or (4) produced factor solutions in which many
items loaded onto more than one factor. Although the
PCA/varimax technique yielded 10 factors (more than one
would hope for), these factors encompassed almost all
items, did not duplicate many items in more than one fac-
tor, and produced factors that made sense. The selection of
items within factors was determined by using a cutoff level
of 0.40 as suggested by Bryant and Yarnold (1995). In addi-
tion, internal consistency was considered a secondary but
relevant consideration. Cronbach’s alpha (Cronbach, 1951)
was used as an index of the degree of internal consistency of
the FaB scale and of each factor subscale. It was expected
that reliability for factors with few variables would be rela-
tively low (Kim & Mueller, 1978). Following completion of
the factor analysis, factor subscales were computed, using
July/August 2003, Volume 57/Number 4
 the item loading coefficients as weightings for use in com- Table 1. Living Situation of Respondents, N = 417
panion analyses. No. (%)

Test–Retest Reliability. The refined scale was field tested Type of dwelling†
Single story house or villa 232 (56)
with a new sample of older people (n = 37) to determine its Ground floor unit 85 (21)
test–retest reliability using intraclass correlation coefficients Multi-story house 41 (10)
Upstairs unit without a lift 30 (7)
(ICC) (2,1) (Shrout & Fleiss, 1979). ICC (2,1) is the Upstairs unit with a lift 26 (6)
appropriate model when generalizing sample results to Community living‡
other judges and respondents. The analysis was done using Community residence 302 (73)
Retirement complex (independent living unit) 109 (26)
the items as the two judging points and the target as the Retirement complex (hostel, common dining area) 4 (1)
people completing the scale. Thus, the ICC is being used to Number of times left home in past month‡
Four or more times a week 268 (65)
tell us the reliability of judgments made for each item at Two or three times a week 102 (25)
baseline and then approximately 2 weeks later. Once a week 25 (6)
Construct Validity. Construct validity offers evidence Once or twice in past month 13 (3)
Never 7 (2)
that a measure reflects the meaning of the construct by †Four missing responses
demonstrating logical relationships between the measure ‡Three missing responses
and other variables (Babbie, 1992). Personal characteristics
were age, gender, type of dwelling, scores on the SF-36 such as climbing one flight of stairs (57%) and walking half
Physical Functioning measure, and number of times left a kilometer (60%). A t-test formula for two sample means
home in the past month. For ordinal data, Spearman rank (Welkowitz, Ewen, & Cohen, 1976) was used to calculate
correlation coefficients were used to estimate associations significant differences between the SF-36 scores for the FaB
between FaB scale scores and age, physical functioning, fre- and comparisons with Australian normative SF-36 data
quency of community outings, and number of falls. For (Australian Bureau of Statistics, 1995). Our respondents’
numeric data, t tests and analysis of variance (ANOVA) perceptions of their physical functioning was similar to the
were used to test relationships between the total FaB scale Australian population norms with no significant differences
scores and the FaB factor subscales with whether people had found. Thirty-four percent (n = 139, 4 missing) of respon-
fallen or not in the past year and gender. Only FaB factor dents reported they had had one or more falls in the past
subscales with modest to good internal consistency were year. This percentage was similar for males and females.
used in this latter analysis and Bonferroni adjustments were Sixty-five people (16%) reported one fall, 37 (9%) report-
made to test for significance when analysis used multiple t ed two falls, 22 (5%) reported three falls, and 14 people
tests. reported four or more falls (3.4%).

Results of Factor Analysis

Results
A total of 422 FaB scales and surveys were returned (76%
response). Four FaB scales were eliminated as they were
incomplete. There were 418 FaB scales used in the analyses
with one of these respondents not completing the demo-
graphic and SF-36 survey.
Study Participants
The age range of respondents was 65 to 98 years with a
mean age of 76.8 years and a majority of women (n = 321,
77%). Most (n = 339, 81%) were born in Australia with an
additional 47 (11.3%) born in Britain and Ireland. As
expected, an overwhelming 97 % (n = 406) spoke English
as the major language at home. Table 1 shows that most
subjects lived in a single-family house in the community
and left their house more than once a week.
The SF-36 scores showed that most people (82%) were
independent in bathing and dressing and over half the
respondents reported that they were not limited in activities
The American Journal of Occupational Therapy
Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930150/ on 07/09/2017 Terms of Use: http://AOTA.org/terms
Following factor analysis, the 45-item version of the FaB
scale was reduced to 30 items. Eight items were dropped
because they were extremely skewed and had poor kurtosis
indicating they were not discriminating between people.
For example, nearly everyone responded “always” to “I
clean up spills as soon as they happen.” Two items were
deleted because they did not contribute anything to the fac-
tors, as they failed to load at, or above, 0.4 on any factor.
These two items were “I have thought about how I would
get up from the floor if I had a fall at home” and “I wait till
the bus stops before I stand up.” As planned, two of the
three items on bifocals were removed (see Design and
Methods) with the item with the most normal distribution
being retained. Three other items were dropped because
they substantially reduced the Cronbach alpha reliability of
the scales to which they belonged. These items were “I
make time to exercise,” “I use a non-slip mat in the bath or
shower,” and “I notice cracked or uneven pathways.” With
varimax rotation a 10-factor solution was obtained from the
383
 final factor analysis of the 30-item scale accounting for Table 2. Dimensions of the FaB
69.9% of the variance. FaB Behavioral Factors
Factor Item Loading
Variance Coefficient
Table 2 lists the factors and the items that were used to
Factor one: Cognitive Adaptations (n = 297) 21.6%
interpret each dimension, and the loading coefficients that, I cross at traffic lights or pedestrian crossings .741
essentially, are measures of centrality of the item to the whenever possible.
factor. When I go outdoors I think about how to move .652
around carefully.
The first factor, or dimension, explains the largest part When I walk outdoors I look ahead for potential .618
of the variance (21.6%) and is composed of Cognitive hazards.
Adaptations, which are behaviors associated with thinking I hold onto a handrail when I climb stairs. .622
When feeling ill take special care of how get up .526
and planning. The second factor, Protective Mobility
and move.
Behaviors, explains 8.6% of the variance and describes When feeling unwell take particular care everyday .488
strategies used when negotiating the environment in a sup- things.
I get help when I need to reach something very .427
portive or protective manner or, conversely, in a way that
high.
the person is not concerned with balance. The third factor Factor two: Protective Mobility (n = 274) 8.6%
describes an Avoidance dimension (7.9% of the variance). I use a walking stick or walking aid when I need it. .709
Here the person is avoiding (or, alternatively, not avoiding) I do things at a slower pace. .701
When I stand up I pause to get my balance. .676
risky situations. The fourth factor describes an Awareness I bend over to reach something only if I have a .599
dimension (5.7% of variance), that is, behaviors associated firm handhold.
with noticing things such as traffic way hazards. Pace, the I avoid ramps and other slopes. .476
When I am feeling unwell I take particular care .425
fifth factor, accounts for 5.6% of the variance. These people doing everyday things.
either avoid doing things quickly or, alternatively, are not Factor three: Avoidance (n = 226) 7.9%
concerned with taking care and cope with going about their I get help when I need to change a light bulb. .808
I use a light if I get up during the night. .687
daily activity at a faster pace. Factor six describes Practical I carry groceries up the stairs only in small .534
Strategies (4.5% of the variance) that often involve antici- amounts.
pating or planning. Factor seven, is Activities That Cause I get help when I need to reach something very .499
high.
Displacement, in particular, going out on windy days (4.5% I adjust the lighting at home to suit my eyesight. .494
of the variance). Factor eight is behaviors associated with I avoid ramps and other slopes. .419
Being Observant (4.1% of the variance). Factor nine Factor four: Awareness (n = 253) 5.7%
I ask my pharmacist or physician questions about the .739
describes behaviors about Coping with Changes in Levels side effects of my medications.
(3.8% of the variance) suggesting the person has strategies I keep shrubbery and plants trimmed back on .737
in place (or, not in place) to cope with higher activity levels. the pathways to my front and back doors.
I notice spills on the floor. .663
And, factor ten, a one-item dimension, Getting to the Phone I clean my spectacles. .607
(3.5% of the variance) is about taking care (or not taking Factor five: Pace (n = 386) 5.6%
care) getting to or reaching for things, such as, the phone. I hurry when I do things. .805
I turn around quickly. .756
Content Validity Factor six: Practical Strategies (n = 304) 4.5%
I avoid walking about in crowded places. .699
The overall content validity of the final 30-item FaB scale I talk with someone I know about things I do .623
that might prevent a fall.
was determined using Waltz and Bausells’s index (1981).
To reach something up high I use the nearest .458
The Content Validity Index was (CVI = 28/30) 0.93. Thus, chair, or whatever furniture is handy, to climb on.
the high level of agreement between judges supports the I notice spills on the floor. -.417
Factor seven: Displacing Activities (n = 385) 4.5%
final version of the FaB scale as having content validity for
I go out on windy days. .854
measuring behaviors associated with falls in older people. Factor eight: Being Observant (n = 378) 4.1%
When I buy shoes I check the soles to see if .762
Reliability they are slippery.
When I walk outdoors I look ahead for potential .410
The 30 item FaB scale had a Cronbach alpha internal con- hazards.
sistency of 0.84. The internal consistencies of the factors Factor nine: Changes in Level (n = 209) 3.8%
were: Cognitive adaptations α = 0.66 (7 items); Protective When wearing bifocals I misjudge a step or do .782
not see a change in floor level.
mobility α = 0.81 (6 items); Avoidance α = 0.66 (6 items); When I am getting down from a ladder or step .578
Awareness α = 0.53 (4 items); Pace α = 0.80 (2 items); stool I think about the bottom rung or step.
Practical strategies α = 0.10 (4 items) (if delete item “notice Factor ten: Getting to the Phone (n = 398) 3.5%
I hurry to answer the phone. .760
spills” α = 0.43); Being observant α = 0.42 (2 items); and
384 July/August 2003, Volume 57/Number 4
Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930150/ on 07/09/2017 Terms of Use: http://AOTA.org/terms
 Changes in level a = 0.22 (2 items). “Displacing Activities” between the SF-36 Physical Functioning scores and the FaB
and “Getting to the Phone” each comprised only one item dimensions, which were strongest for Practical Strategies,
and so α was not calculated. Reliability analysis demon- Protective Mobility, Cognitive Adaptations, and Avoidance.
strated that “Notice Spills” did not contribute to the relia- Frequency of leaving home was also negatively correlated
bility of the Practical Strategies dimension, however, it was with the FaB factor subscales. The total FaB mean score
a contributing item to the Awareness dimension. Small showed similar trends. This means that the younger and
numbers of items in some factors means that it was difficult more mobile the person, the less the person demonstrated
to obtain a reasonable internal consistency. safety behaviors.
The test–retest reliability study included 37 FaB scales Four out of the 10 subscales were included in further
that were returned from a total of 40 distributed. The aver- testing. There were statistically significant differences
age age of respondents was 80.03 (69 to 93) years. The aver- between scores on two of these FaB factor scales comparing
age number of days between the test and retest was 16.25 those who recalled that they had fallen (n = 139) with those
(10 to 27) days. The ICC for the total FaB mean scores who did not recall a fall (n = 275) in the past year. These
between the first data collection and the retest was 0.94 (p were Protective Mobility (fallen M = 0.55, SD = 0.16; not
< .01). The consistency of judgments over the 2-week peri- fallen M = 0.49, SD = .18) t(412) = 3.17, p < .01 and
od for each item ranged from Feeling unwell take care 0.61 Cognitive Adaptations (fallen M = 0.75, SD = 0.14; not
(p < .01) to Use walking aid when needed 0.98 (p < .01). fallen M = 0.71, SD = 0.15) t(412) = 2.62, p < .01.
Most items had an ICC above 0.75 with 17 items 0.85 or Following Bonferroni adjustment, Avoidance failed to reach
higher. The test–retest reliability of the Factor scales ranged statistical significance (fallen M = 0.52, SD = 0.12; not fall-
from Changes in Level 0.78 (p < .01) to Protective Mobility en M = 0.49, SD = 0.15) t(412) = 2.28, p = 0.02. There was
0.96 (p < .01). no difference in the way that the dimension Pace was used
between those who fell in the past year (M = 0.37, SD =
Relationship of FaB to Falls 0.11) and those who did not fall (M = 0.36, SD = 0.12)
A statistically significant, though not strong, difference was t(412)=0.92, p = 0.36.
found in FaB mean scores between those who reported a fall Females (n = 321; M = 3.02, SD = 0.47) used safer
in the previous year (M = 3.05, SD = 0.47) and those who behaviors than males (n = 96; M = 2.81, SD = 0.50)
did not report a fall (M = 2.93, SD = 0.49), t(412) = 2.48, although this was not a strong association, t(415) = 3.87, p
p < .05. The 95% confidence interval of the difference < .01. The largest gender difference for the subdimensions
between means (0.12) was 0.02–0.22. was for Avoidance behaviors, t(415) = 6.37, p < .01, which
There were positive though weak-to-moderate correla- were more common among females (M = 0.53, SD = 0.13)
tions between increasing age and most of the dimensions of than males (M = 0.42, SD = 0.15). There was no difference
the FaB scale (see Table 3). There were negative correlations between the male and female mean scores of the Pace
dimension (females M = 0.36, SD = 0.11; males M = 0.38,
Table 3. Rank-Order Correlations Between FaB Scores and SD = 0.12) not significant at t(415)=1.22 or the Protective
Risk Factors for Falls: Age, SF-36 Scores, and Number of Times
Left Home Mobility dimension (females M = 0.51, SD = 0.17; males
FaB scales Age category SF-36 Physical Times left M = 0.49, SD = 0.18) not significant at t(415) = 0.99.
Functioning home
rs rs rs
(n = 417) (n = 395) (n = 415) Discussion
Cognitive Adaptations .428** -.581** -.402**
Protective Mobility .440* -.671** -.506** The development of the Falls Behavioral Scale for Older
Avoidance .381** -.581** -.370** People has provided a reliable tool (internal consistency ∝ =
Awareness .105* -.077 -.045
Pace .353** -.493** -.446** 0.84; test–retest reliability ICC = 0.94) with high content
Practical Strategies .402** -.674** -.452** validity (CVI = 0.93) for determining the presence or
Displacing Activities .234** -.310** -386**
(n = 384) (n = 366) (n = 382)
absence of protective behaviors. The factor analysis refined
Being Observant .260** -.381** -.233** the tool and isolated 10 dimensions that contribute to
Changes in Level -.131** -.212** .231** understanding the kinds of actions and behaviors associat-
Getting to Phone .128* -.270** -.227**
(n = 397) (n = 378) (n = 395) ed with older people making adaptations to enhance safety
Total FaB mean score .464** -.684** -.514** and to protect themselves from falling. The relationship
*Significant at the .05 level. between the FaB and personal characteristics supports the
**Significant at the .01 level.
FaB: Falls Behavior Scale for Older People construct validity of the FaB for use with older people. The
SF-36: SF-36 Health Survey (Ware, Snow, Kosinski, & Gandek, 1993) directions of the measured relationships were what would
The American Journal of Occupational Therapy 385
Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930150/ on 07/09/2017 Terms of Use: http://AOTA.org/terms
 be expected, with more protective behaviors associated with
increasing age, reduced mobility, and reduced frequency
that the older person left home, which are also known risk
factors for falls (e.g., Tinetti et al., 1988).
The FaB scale shows potential for measuring behav-
ioral change. In this study the FaB scale suggested that peo-
ple with a history of falling appear to make adaptations and
use safer practices than those who do not have a history of
falling. The differences found were statistically significant,
though not strong, between those that reported a fall in the
previous year and those who did not report a fall. The rea-
sons for making the kinds of adaptations the FaB scale
measures may be explained by a number of factors includ-
ing falling. People make adaptations for various reasons,
for example, functional visual loss, or conditions such as
stroke that decrease balance and strength. However, as sig-
nificant differences were found this provides a benchmark
to compare with future investigations. These results suggest
that the FaB is a useful scale that could be used to measure
the effect of a program that aims to reduce risky behaviors
and enhance safe adaptations associated with falls preven-
tion.
The scale could also be used in clinical practice in a
home visit situation as part of an evaluation of the person
and their risk. As with an environmental assessment
(Clemson, 1997), occupational therapists are familiar with
placing in context and taking into account the person’s
physical, functional, and motivational status and their envi-
ronment when interpreting assessment results and making
decisions about the person’s degree of risk. The FaB can pro-
vide a profile of the person’s perception of the kinds of
behaviors he or she uses.
Building a Perspective of Behavioral Factors and Falls
Models of behavioral change rely on being able to define the
kinds of everyday behaviors and actions targeted for change
(Maddux & DuCharme, 1997). The dimensions defined
by the factor analysis contribute to profiling different types
of fallers or persons who may be at different stages of adap-
tation, for example, those people who negotiate their envi-
ronment in a protective manner compared with those peo-
ple who have no concerns.
Many of the FaB dimensions, and in particular
Cognitive Adaptations and Awareness, clearly involve cog-
nitive elements such as thinking, planning, recognition, and
anticipation. In-depth qualitative data from fall reenact-
ments by Connell and Wolf (1997) and Clemson et al. (in
press) produced thematic scenarios demonstrating how
behavioral and situational factors contribute to falls. Such
factors include habitual actions within familiar environ-
ments despite deteriorating physical functioning and
386
Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930150/ on 07/09/2017 Terms of Use: http://AOTA.org/terms
reduced awareness of surroundings. These studies suggest
that prevention can include education to increase sensitivi-
ty to one’s immediate environment, to raise awareness of
functional vision, strength, and balance and the interaction
of these abilities with environmental usage, and to recognize
everyday habits that contribute to fall risk. A tool that is
able to identify the older person’s awareness of and practice
of protective behaviors related to falling could be useful to
occupational therapists.
A factor such as Cognitive Adaptations has a strong
community mobility emphasis with many items related to
safely getting about in the local community. About 40% of
falls occur away from home (e.g., Reinsch, Macrae,
Lachenbruch, & Tobis, 1992), yet community mobility is
not a usual component of falls prevention programs.
There were similar gender responses for Protective
mobility and Pace suggesting adaptations to reduced phys-
ical function are similar for males and females. Our expe-
rience working with older people in our community-based
falls prevention program (Clemson et al., 2003) and sup-
ported by other data (Clemson et al., in press), is that both
males and females attribute hurrying as a contributor to
falls and “slowing down” as a strategy or phenomenon they
consciously utilize to cope with reduced physical function-
ing, and in some cases, with falling. We found the most
marked gender difference on the FaB scale was that women
favored Avoidance behaviors more than men. Waldron
(1988) has demonstrated how factors such as women’s par-
enting responsibilities compared with men’s greater
assertiveness and rebelliousness has resulted in a greater
tendency for females to engage in avoidance behaviors,
such as avoiding smoking. Some situations that are avoid-
ed could also be explained in terms of traditional gender
roles, for example, changing a light bulb is one potential
fall situation that some women tend to avoid. We have
found that some men admit that it is hard to cease climb-
ing ladders as this is what they have always done and view
it as associated with their home maintainer role. The safe
performance of certain physical activities rather than
avoidance is consistent with older people remaining as
independent as possible (Speechley & Tinetti, 1991).
Nonetheless, our results show that for some older people,
and to a degree more often women, avoidance is a strategy
used to avoid risk.
The FaB assessment was designed to rely on the per-
son’s perceptions of his or her own behaviors. Several items
were removed during the tool development because most
people answered in the same way, one example being that
most people perceived that they did not leave objects lying
on the floor in and about walkways. It may be that, as noted
by Nevitt et al. (1989) when determining reliability of a
July/August 2003, Volume 57/Number 4
 take home hazard checklist, that we tend not to “see” clut-
ter within our own environment. Although the entire FaB
scale was reliable, small numbers of items in factors meant
that internal consistency was low on some of the subscales.
It seems that we may have more confidence in the total scale
and those individual subscales that were most reliable, than
all of the individual subscales when using them as a
measure.
The FaB scale has potential utility for occupational
therapists as an assessment and goal-setting tool in clinical
practice. It can provide a profile of the range of safety strate-
gies and the kinds of restrictive behaviors a person is using.
The scale can be used as a prompt to discuss behavioral fac-
tors and falls and as an aide in reflective learning. For exam-
ple, if mailed to a client prior to a home visit the scale can
be used to raise awareness of a broad therapist focus on
behavioral as well as environmental change.
Conclusion
These results support the presence of a number of underly-
ing dimensions that together constitute what we broadly
classified as behavioral factors related to falls prevention in
older age. These dimensions appear to be strongest for
Cognitive Adaptations, Protective Mobility, Avoidance, and
Pace. Other dimensions isolated were Awareness, Practical
Strategies, Displacing Activities, Being Observant, Changes
in Level, and Getting to the Phone. The FaB scale appears
to be a quick, useful, reliable, and valid tool for measuring
behaviors associated with falls prevention in older age. It has
potential to be used as an epidemiological assessment tool
to show a difference in the participants’ behaviors compared
to controls. It could also be used pre- and postfalls preven-
tion intervention to reflect change in participant’s use of
safe behavioral strategies. In addition, the FaB scale could
be a useful tool for occupational therapists and other health
professionals to use with individuals to assist in assessment
and as a prompt to discussion in goal setting for falls
prevention.▲
Acknowledgments
The expert reviewers were Sue Carter, Robert Cumming,
Gina Frampton, Vera Frazer, Mrs. Heard, Lynette
Mackenzie, William Mann, Megan Swann, and Robyn
Twible. Those who assisted with field testing the scale
included Eastwood Christian Community Aid, Mercy
Family Centre at Hornsby, Eastwood Senior Citizens,
Auburn Aged Care Centre, Jean Hill from Alan Walker
Village, and Muriel Clemson.
The American Journal of Occupational Therapy
Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930150/ on 07/09/2017 Terms of Use: http://AOTA.org/terms
References
Australian Bureau of Statistics. (1995). National health survey: SF-
36 population norms, Australia. (ABS Cat. No. 4399.0).
Canberra, Australia.
Babbie, E. (1992). The practice of social research (6th ed.).
Belmont, CA: Wadsworth.
Bryant, F. B., & Yarnold, P. R. (1995). Principal-components
analysis and exploratory and confirmatory factor analysis. In
L. G. Grimm & P. R. Yarnold (Eds.), Reading and under-
standing multivariate statistics (pp. 99–136). Washington,
DC: American Psychological Association.
Clemson, L. (1997). Home fall hazards and the Westmead Home
Safety Assessment. West Brunswick, Australia: Co-ordinates
Publications.
Clemson, L., Cusick, A., & Fozzard, C. (1999). Managing risk
and exerting control: Determining follow through with falls
prevention. Disability and Rehabilitation, 13(12), 531–541.
Clemson, L., Fitzgerald, M. H., & Heard, R. (1999). Content
validity of an assessment tool to identify home fall hazards:
The Westmead Home Safety Assessment. British Journal of
Occupational Therapy, 62(4), 171–179.
Clemson, L., Manor, D., & Fitzgerald, M. H. (in press).
Behavioral factors contributing to older persons falling in
public places. Occupational Therapy Journal of Research:
Occupation, Participation and Health.
Clemson, L., Swann, M., Twible, R., Cumming, R. G., Kendig,
H., & Taylor, K. (2003). Stepping out: Building confidence
and reducing falls. A community based program for older peo-
ple. Lidcombe, New South Wales, Australia: The University
of Sydney.
Close, J., Ellis, M., Hooper, R., Glucksman, E., Jackson, S., &
Swift, C. (1999). Prevention of falls in the elderly trial
(PROFET): A randomised controlled trial. Lancet, 353,
93–97.
Cole, H., Berger, P., & Garrity, T. (1988). Analogues between
medical and industrial safety research on compliance
behaviour. In D. Gochman (Ed.), Health Behaviour:
Emerging Research Perspectives (pp. 337–353). New York:
Plenum.
Comrey, A. L., & Lee, H. B. (1992). A first course in factor anal-
ysis (2nd ed.). Hillsdale, NJ: Erlbaum.
Connell, B., & Wolf, S. (1997). Environmental and behavioural
circumstances associated with falls at home among healthy
elderly individuals. Archives of Physical & Medical
Rehabilitation, 78(2), 179–186.
Cronbach, L. J. (1951). Coefficient alpha and the internal struc-
ture of tests. Psychometrika, 16, 297–334.
Cumming, R. G., Thomas, M., Szonyi, G., Frampton, G.,
Salkeld, G., & Clemson, L. (2001). Adherence to recom-
mendations by an occupational therapist for home modifica-
tions for fall prevention. American Journal of Occupational
Therapy, 55, 641–648.
Cumming, R. G., Thomas, M., Szonyi, G., Salkeld, G., O’Neill,
E., Westbury, C., et al. (1999). Home visits by an occupa-
tional therapist for assessment and modification of environ-
mental hazards: A randomized trial of falls prevention.
Journal of American Geriatrics Society, 47(12), 1397–1402.
387
 Cummings, S. R., & Nevitt, M. C. (1989). A hypothesis: The
causes of hip fractures. Journal of Gerontology, 44(4),
M107–M111.
Ebel, R. L. (1979). Essentials of educational measurement (3rd ed.).
Englewood Cliffs, NJ: Prentice–Hall.
Gochman, D. S. (Ed.). (1997). Handbook of health behaviour
research 1. Personal and social determinants. New York:
Plenum Press.
Hill, K., Schwarz, J., Flicker, L., & Carroll, S. (1999). Falls
among healthy community dwelling older women: A
prospective study of frequency, circumstances, consequences
and prediction accuracy. Australian and New Zealand Journal
of Public Health, 23, 41–48.
Jackson, J. (1996). Living a meaningful existence in old age. In R.
Zemke & F. Clark (Eds.), Occupational science. The evolving
discipline (pp. 339–361). Philadelphia: F.A. Davis.
Kaiser, M. (1958). The varimax criterion for analytic rotation in
factor analysis. Psychometrika, 23, 187–200.
Kim, J., & Mueller, C. (1978). Factor analysis. Statistical methods
and practical issues. Beverly Hills, CA: Sage.
Kline, P. (1994). Easy guide to factor analysis. New York:
Routledge.
Lynn, M. R. (1986). Determination and quantification of con-
tent validity. Nursing Research, 35(6), 382–385.
Lyons, R. A., Perry, H. M., & Littlepage, B. N. C. (1994).
Evidence for the validity of the Short-form 36 questionnaires
(SF-36) in an elderly population. Age and Ageing, 23,
182–184.
Maddux, J. E., & DuCharme, K. A. (1997). Behavioural inten-
tions in theories of health behaviour. In D. S. Gochman
(Ed.), Handbook of health behavior research 1. Personal and
Social Determinants (pp. 133–151). New York: Plenum.
Nevitt, M. C., Cummings, S. R., Kidd, S., & Black, D. (1989).
Risk factors for recurrent nonsyncopal falls: A prospective
study. JAMA, 261(18), 2663–2668.
Reinsch, S., Macrae, P., Lachenbruch, P. A., & Tobis, J. S. (1992).
Why do healthy older adults fall? Behavioural and environ-
mental risks. Physical & Occupational Therapy in Geriatrics,
11(1), 1–15.
388
Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930150/ on 07/09/2017 Terms of Use: http://AOTA.org/terms
Ronis, D. L., Yates, J. F., & Kirscht, J. P. (1989). Attitudes, deci-
sions, and habits as determinants of repeated behavior. In A.
R. Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds.),
Attitude, structure and function (pp. 213–239). New York:
Erlbaum.
Shiffrin, R. M., & Schneider, W. (1977). Controlled and auto-
matic human information processing: II. Perceptual learn-
ing, automatic attending, and a general theory. Psychological
Review, 84, 127–190.
Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses
in assessing rater reliability. Psychological Bulletin, 86(2),
420–428.
Skakun, E. N., & Kling, S. (1980). Comparability of methods for
setting standards. Journal of Educational Measurement, 17,
229–235.
Speechley, M., & Tinetti, M. (1991). Falls and injuries in frail and
vigorous community elderly persons. Journal of American
Geriatric Society, 39, 46–52.
SPSS, Inc. (1999). SPSS base 10.0 applications guide. Chicago:
Author.
Thurstone, L. L. (1947). Multiple factor analysis. Chicago:
University of Chicago Press.
Tinetti, M. E., Speechley, M., & Ginter, S. F. (1988). Risk factors
for falls among elderly persons living in the community. New
England Journal of Medicine, 319(26), 1701–1707.
Tinetti, M. E., & Williams, C. S. (1997). Falls, injuries due to
falls, and the risk of admission to a nursing home. New
England Journal of Medicine, 337(18), 1279–1284.
Waldron, I. (1988). Gender and health related behaviour. In D.
Gochman (Ed.), Health Behaviour: Emerging Research
Perspectives (pp. 193–208). New York: Plenum.
Waltz, C., & Bausell, R. B. (1981). Nursing research: Design,
statistics & computer analysis. Philadelphia: F. A. Davis.
Ware, J. E., Snow, K. K., Kosinski, M., & Gandek, B. (1993). SF-
36 Health Survey manual & interpretation guide. Boston:
Health Institute, New England Medical Center.
Welkowitz, J., Ewen, R. B., & Cohen, J. (1976). Introductory
statistics for the behavioural sciences. New York: Academic
Press.
July/August 2003, Volume 57/Number 4